174 NATURAL SCIENCE. Sept.. 



the New and Old World. These families must possess some 

 characters which confer upon them the power of assuming an 

 epiphytic mode of existence. 



The first indispensable feature is that the seed shall be able to 

 reach the tree on which it is to grow. The only two practicable 

 methods of seed-dissemination are by the wind or by the agency of 

 animals. This fact has been well illustrated by the observations of 

 Loew (4) on the means of distribution of Epiphytes found on willow- 

 tops in Germany. A walk along the banks of an English river, 

 shaded by pollard-willows, reveals the fact that there is a rich epi- 

 phytic flora on these trees. Among them are not only algae, and mosses, 

 but also numerous herbs, and large rose and blackberry bushes. Willis 

 and Burkill (5) found that, near Cambridge, 44-6 per cent, of these Epi- 

 phytes had fleshy fruits ; i6'4 per cent, had burrs ; 25-1 per cent, had 

 winged or feathery fruits ; io-6 per cent, had very light seeds ; while 

 in only 2*9 per cent, was the means of distribution feeble or obscure. 



So, in the tropics, the seeds of Epiphytes may be all ranged 

 under three main classes : — 



(i.) Those enveloped in a fleshy covering (many bromeliads) ; 

 (ii.) Seeds (or spores) very light and easily carried by the wind 

 (Dendrobiumaitemiatmn, the seed of which weighs -00000565 

 grammes) ; 

 (iii.) Seeds light and with a mechanism for floating in the air. 

 ("Wings" in Rhododendvon pendtdimi, in which the seed 

 weighs -000028 grammes : tufts of hairs in Asclepiadaceae : 

 air-containing spaces in the envelope, ^schynanthus 

 pulchra.) 



In all cases of these regular tropical Epiphytes the seeds are 

 small, so that they can easily become lodged in minute crevices 

 on the surface of the host. Those characters are not adaptive. 

 The seeds must have possessed them to a large extent when 

 the plants were terrestrial ; otherwise the latter could have never 

 become Epiphytes. The structure of their fruits and seeds 

 explains why certain families should not have Epiphytes among 

 them — thus the heavy-seeded Leguminosae. On the other hand, in 

 certain families only those types which possessed appropriate seeds 

 have become Epiphytes. For example, the Loganiaceae frequently 

 possess fleshy fruits, but the seeds are large excepting in Fagraa, which 

 accordingly has epiphytic species. 



Even when the seed has reached the surface it must rapidly 

 make arrangements for the firm attachment of the embryo. Only a 

 few observations exist on the germination of Epiphytes {Clusia, Myrme- 

 codia, /Eschynantkus, Tceniophyllum). In all these the primary root soon 

 ceases to grow, and there is a precocious development of an extensive 

 system of adventitious roots. In ^schynanthus pulchra (3) the radicle 

 remains short, the inferior end of the embryo becomes flattened, and 



